Netflix expanded its EBIT margin by approximately 10.5 percentage points while increasing cash content spending only 2% annually. That is what valuable scale looks like: spending slows, margins expand, and more earnings become cash.
Netflix Turned Spending Restraint Into Faster Earnings Growth
Since 2021, Netflix’s EBIT margin has risen from 21% to approximately 31.5%, even as cash content spending grew at only a 2% annual rate.
Over five years, revenue grew 12% annually, while operating profit grew 21% and EPS grew 27%. Netflix now converts approximately 90% of earnings into free cash flow, which is primarily used for repurchases.
The widening gap between revenue and earnings growth suggests that Netflix is extracting more economic value from each additional dollar of revenue. Repurchases then extend those operating gains into per-share economics.
Better Economics Make the Valuation Reset Matter
Bill Ackman returned to Netflix after an approximately 50% decline from its June 2025 high of $134. The decline reduced its forward earnings multiple from more than 40 times to 21 times.
The valuation reset matters, but the measurable improvement in Netflix’s economics matters more.
Advertising provides another potential growth engine, with revenue approaching $3 billion this year. Ackman estimates that earnings could compound at close to 20% annually. That remains an investor judgment, though it rests on an operating model whose margin and cash flow progression can now be measured.
Netflix also lost its Warner Bros. Discovery bid in February 2026 and collected a $2.8 billion termination fee, removing one acquisition uncertainty.
SK Hynix Is Trying to Make Memory Less Commodity-Like
SK Hynix held 58% of the first-quarter HBM market, compared with 21% each for Samsung and Micron. In July, it reported 10 long-term customer agreements.
Nvidia also agreed to co-develop next-generation memory within a $500 billion agreement with SK Group. Custom HBM products, including HBM5 tailored to processors from customers such as Nvidia and Google, could deepen these relationships.
Multi-year agreements, customer-specific products, and co-development can make capacity harder to substitute. That is the strongest argument that this memory cycle may be structurally different. But more durable returns have not yet been proved.
Capital Spending Threatens the Scarcity Supporting Returns
SK Hynix is investing $720 billion in what it says will become the world’s largest network of memory factories, alongside a $4 billion packaging facility in Indiana. Micron is spending $50 billion on two Boise fabs and building a $100 billion New York campus.
Applied Materials reported strong DRAM and advanced-packaging demand and raised its calendar 2026 semiconductor-systems expectations. Yet its shares initially fell about 4.6% after hours despite results above estimates, suggesting that strength is already embedded in expectations.
SanDisk’s target of about 80% non-GAAP gross margin during 2028 through 2030 makes the optimism surrounding memory economics unusually explicit.
The tension is straightforward. SK Hynix may be building contractual and technical advantages that make HBM less commodity-like, while the industry commits immense amounts of capital to new capacity.
Scarcity supports attractive economics. Capacity expansion attacks scarcity.
Netflix shows what valuable scale looks like after spending slows. SK Hynix still must prove that extraordinary spending will preserve scarcity rather than destroy it.